With the concept of knowledge graph proposed by Google in 2012,it has gradually become a research hotspot in the field of artificial intelligence and played a role in applications such as information retrieval,question answering,and decision analysis.While the knowledge graph shows its potential in various fields,it is easy to find that there is no mature knowledge graph construction platform currently.Therefore,it is essential to research the knowledge graph construction system to meet the application needs of different industries.This paper focuses on the construction of the knowledge graph.Firstly,it introduces the current mainstream general knowledge graphs and domain knowledge graphs and describes the differences between the two in the construction process.Then,it discusses the problems and challenges in the construction of the knowledge graph according to various types.To address the above-mentioned issues and challenges,it describes the five-level solution methods and strategies of knowledge extraction,knowledge representation,knowledge fusion,knowledge reasoning,and knowledge storage in the current graph construction process.Finally,it discusses the possible directions for future research on the knowledge graph and its application.

How to accurately model user preferences based on existing user behavior and auxiliary information is of great important.Among all kinds of auxiliary information,Knowledge Graph (KG) as a new type of auxiliary information,its nodes and edges contain rich structural information and semantic information,and attracts a growing researchers' attention in recent years.Plenty of studies show that the introduction of KG in personalized recommendation can effectively improve the performance of recommendation,and enhance the rationality and interpretability of recommendation.However,the existing methods either explore the independent meta-paths for user-item pairs over KG,or adopt graph representation learning on whole KG to obtain representations for users and items separately.Although both have achieved certain effects,the former fails to fully capture the structural information of user-item pairs in KG,while the latter ignores the mutual effect between target user and item during the embedding propagation.In order to make up for the shortcomings of the above methods,this paper proposes a new model named User-end and Item-end Knowledge Graph (UIKG),which can effectively capture the correlation between users' personalizedpreferences and items by mining the associated attribute information in their respective KG.Specifically,we learn the user representation vectors from the user KG,and then introduce the user representation vectors into the item KG based on the method of graph convolution neural network to jointly learn the item representation vectors,so as to realize the seamless unity of the user KG and the item KG.Finally,we predict the user preference probability of the item through MLP.Experimental results on open datasets show that,compared with the baseline method,UIKG improves by 2.5%~13.6% on Recall@K index,and 0.4%~5.8% on AUC and F1 indexes.

Knowledge graph is a structured representation of real-world triples.Typically,triples are represented in the form of head entity,relationship entity and tail entity.Aiming at the data sparse problem widely existing in knowledge graph,this paper proposes a knowledge graph completion method using quaternions as relational rotation.In this paper,we model entities and relations in the expressive hyper-complex representations for link prediction.This hyper-complex embedding is used to represent entities,and relations are modelled as rotations in quaternion space.Specifically,we define each relation as a rotation from the head entity to the tail entity in the hyper-complex space,which could be used to infer and model diverse relation patterns,including symmetry/anti-symmetry,reversal and combination.In the experiment,the public datasets WN18RR and FB15K-237 are used for the related link prediction experiment.Experimental results show that on the WN18RR dataset,its mean reciprocal rank (MRR) is 4.6% higher than RotatE,and its Hit@10 is 1.7% higher than RotatE.On the FB15K-237 dataset,its MRR is 5.6% higher than RotatE,its Hit@3 is 1.4% higher than RotatE.Experiments show that the knowledge graph completion method using quaternions as relational rotation can effectively improve the prediction accuracy of triples.

Service discovery aims to solve the problem of service information explosion,find and locate services that meet the needs of service requesters.Since the service description information is mainly composed of short text with noise and has the feature of sparse semantics,it is difficult to extract the implicit context information of the service description document.In addition,the traditional service discovery method directly obtains the characteristic representation of the service.According to the cosine similarity to calculate the similarity,the used measurement function is not in line with human perception.In response to the above two problems,this paper proposes a service discovery framework (KSN) based on knowledge graphs and neural similar networks.It uses the knowledge graph to connect the entities in the service description and specifications to obtain rich external information,thereby enhancing the semantic information of the service description.And it uses convolutional neural network (CNN) to extract the feature vector of the service as the input of the neural similarity network.The neural similarity network will learn a similarity function to calculate the similarity between the service and the request to support the service discovery process.A large number of experiments on real service data sets crawled by ProgrammableWeb show that KSN is superior toexisting Web service discovery methods in terms of multiple evaluation metrics.

For personalized recommendation,common recommendation algorithms include content recommendation,Item CF and User CF.However,most of these algorithms and their improved algorithms tend to focus on users' explicit feedback (tags,ra-tings,etc.) or rating data,and lack the use of multi-dimensional user behavior and behavior order,resulting in low recommendation accuracy and cold start problems.In order to improve the recommendation accuracy,a collaborative filtering recommendation algorithm based on knowledge graph (BR-CF) is proposed.Firstly,according to the user behavior data,behavior graph and behavior route are created considering the behavior order,and then the vectorization technology (Keras Tokenizer) is used.Finally,the similarity between multi-dimensional behavior route vectors is calculated,and the route collaborative filtering recommendation is carried out for each dimension.On this basis,two improved algorithms combining BR-CF and Item CF are proposed.The expe-rimental results show that the BR-CF algorithm can recommend effectively in multiple dimensions on the user behavior dataset of Ali Tianchi,realize the full utilization of data and the diversity of recommendation,and the improved algorithm can improve the recommendation performance of Item CF.

Recommendation systems alleviate the problem of information overload caused by the rapid increase of data on the Internet.But traditional recommendation systems are not accurate enough due to data sparsity and cold start.Therefore,a novel recommendation algorithm based on knowledge graph and tag-aware (KGTA) is proposed.First,tags of items and users are used to capture low-order and high-order features through knowledge graph representation learning.The semantic information of entities and relationships in two knowledge graphs is embedded into a low-dimension vector space to obtain the unified representation of items and users.Then,deep neural networks and recurrent neural networks combining attention mechanism are respectively utilized to extract the latent features of items and users.Finally,ratings are predicted on the basis of latent features.KGTA not only takes relationship information and semantic information of knowledge graph and tags into consideration,but also learns latent features of items and users through deep structures.Experimental results on MovieLens datasets illustrate that the proposed algorithm performs better in rating prediction and improves the accuracy of recommendation.

For the problem of data sparsity in recommendation systems,traditional collaborative filtering methods fail to capture the correlations between auxiliary information,which decreases the accuracy of recommendation.In this paper,we propose the KGE-CF model by introducing the knowledge graph as auxiliary information,and utilizing the multi-source structured data in knowledge graph to effectively alleviate data sparsity.KGE-CF integrates multi-layer perceptron to capture the high-order nonli-near features to learn deeper interaction information between users and items,which can improve the quality of recommendation.Specifically,we first map the user's historical interaction items with the corresponding entities in the knowledge graph,and use the translation model of the knowledge graph for training,through that we can get the entity embedding vector and the relation embedding vector,moreover the model can learn a richer user vector by interest disseminating.Then,after concatenating the obtained user vector and the item vector,we send it into the multi-layer perceptron to capture the high-order feature information between the user and the item.Finally,a sigmoid function is used to obtain the user's preference probability for candidate items.The experimental results on the real-world datasets prove that the proposed KGE-CF model in this paper has achieved the best recommendation performance than state-of-art methods.

With the arrival of the big data era,the relationship that needs to be processed in various industries has increased exponentially,and there is an urgent need for a data model that supports the ability to express massive complex relationship,that is,domain knowledge graph.Although the domain knowledge graph has shown great potential,it is not difficult to find that there is still a lack of mature construction technologies and platforms.It still remains an important challenge to construct domain know-ledge graph rapidly.After the systematic study of domain knowledge graph,a method is proposed to construct domain knowledge graph based on property graph model.Concretely,for structured and semi-structured data stored in a variety of databases,the method completes the construction of the high-quality graph model by graph database data communication protocol,multiple configuration methods of entity and relation schema,etc.Then,the data from the original database is extracted,transformed and loa-ded into the property graph database HugeGraph,completing the construction of domain knowledge graph.Finally,experiments on multiple datasets and test results of Gremlin statement show that the proposed method is complete and reliable.

Due to the rapid updating of computer knowledge with many ambiguities,it is difficult for students to seek reasonable solutions for independent innovation.As an auxiliary innovation tool,intelligent question-answering system can help students to grasp the frontier of subject development,find out solutions for problems faster and precisely.In this paper,a knowledge graph of scientific research is constructed based on a large-scale database of scientific and technological documents,which realizes an intelligent question answering system for assisting students in innovation.In order toreduce the influence of noisy entities on query questions,this paper proposes an auxiliary task enhanced intent information for question answering in computer domain(ATEI-QA).Compared with the traditional method,this method can extract the question intention information more accurately and further reduce the influence of noisy entity on intention recognition.Additionally,we conduct a series of experimental studies on computer and common datasets,and compare with three mainstream methods.Finally,experimental results demonstrate that our model achieves significant improvements against with three baselines,improving MAP and MRR scores by average of 3.27%,1.72% in the computer dataset and 4.37%,2.81% in the common dataset respectively.

Existing knowledge graph completion models learn a single static feature representation for entities and relationships by integrating multi-source information.But they can't represent the subtle meaning and dynamic attributes of entities and relationships that appear in different contexts.That is,entities and relationships will show different attributes,because they have different roles and meanings when they are involved in different triples.To solve above problems,an adaptive attention network for knowledge graph completion is proposed,which uses adaptive attention to model the contribution of each task-specified feature dimension,and generates dynamic and variable embedding representations for target entities and relationships.Specifically,the proposed model defines the neighbor encoder and the path aggregator to process two structures in the entity neighborhood subgraph,adaptively learn the attention weights to capture the most logically related features of the task,and to give the entities and relationships with fine-grained semantics in line with the current task.Experimental results in link prediction task show that,the MeanRank of the proposed model on FB15K-237 dataset is 6.9% lower than PathCon,and Hits@1 is 2.3% higher than PathCon.For the sparse datasets NELL-995 and DDB14,its Hits@1 reaches 87.9% and 98% respectively.Therefore,it proves that the introduction of adaptive attention mechanism can effectively extract the dynamic attributes of entities and relationships to generate a more comprehensive embedding representation,and improves the accuracy of knowledge graph completion.

The recommendation system based on graph neural network is the current research hotspot of data mining applications.The recommendation performance can be improved by combining the graph neural network on the heterogeneous information network(HIN).However,the existing HIN-based recommendation methods often have problems that cannot effectively explain the results of high-level modeling,and manual design of meta-paths requires knowledge of related domains.Therefore,this paper combines the idea of hierarchical granulation andproposes a heterogeneous recommendation method(HKR) based on knowledge graphs.The local context and non-local context are hierarchically granulated,and the coarse-grained representation of user characteristics is learned separately.Then based on the gating mechanism, combining local and non-local attribute node embedding,learning the potential features between users and items,and finally fusing fine-grained features for recommendation.The real experimental results show that the performance of the proposed method is better than the current graph neural network recommendation method based on knowledge graph in many aspects.

As a structured form of human knowledge,knowledge graphs have played a great supportive role in supporting the semantic intercommunication of massive,multi-source,heterogeneous data,and effectively support tasks such as data analysis,attracting the attention of academia and industry.At present,most knowledge graphs are constructed based on non-real-time static data,without considering the temporal characteristics of entities and relationships.However,data in application scenarios such as social network communication,financial trade,and epidemic spreading network are highly dynamic and exhibit complex temporal properties.How to use time series data to build knowledge graphs and effectively model them is a challenging problem.Recently,numerous studies use temporal information in time series data to enrich the characteristics of knowledge graphs,endowing know-ledge graphs with dynamic features,expanding fact triples into quadruple representation(head entity,relationship,tail entity,time).The knowledge graph which utilizes time-related quadruples to represent knowledge are collectively referred to as temporal knowledge graph.This paper summarizes the research work of temporal knowledge graph representation learning by sorting out and analyzing the corresponding literature.Specifically,it first briefly introduce the background and definition of temporal know-ledge graph.Next,it summarizes the advantages of the temporal knowledge graph representation learning method compared with the traditional knowledge graph representation learning method.Then it elaborates on the recent method of temporal knowledge graph representation learning from the perspective of the method modeling facts,introduces the dataset used by the above method and summarizes the main challenges of this technology.Finally,the future research direction is prospected.

Question answering over knowledge graph(KG-QA) systems map the natural language question to the 〈subject,predicate,object〉 triple in the knowledge graph(KG) by semantic analysis of the given question,and infer the triple to get the answer of the question.Due to the diversity of natural languages,a question may be expressed in multiple forms but the triples in KGs are structured data in a standard form.It is challenging to map questions to triples in KGs.This paper proposes a novel Key-Value relational memory network,starting from the perspective of KGs,and focusing on the relationship between the candidate answer knowledge and the relationship between the knowledge in KGs and the question representations.In addition,the attention mechanism is applied in the proposed model,so that it has better interpretability than other baseline models.We evaluate the method on WebQuestions benchmark.Experiment results show that,compared with the best methods based on information extraction,the F1 value of the proposed method increases by 5.9% and is slightly higher than that of the optimal methods based on semantic analysis,which verifies the effectiveness of the proposed method.

Influence calculation and analysis are widely used in social networks,web page importance evaluation and other fields.There is still a lack of effective and universal solution for the multi-level influence calculation with inheritance chain and time span factors.At the same time,the calculation of maximizing the propagation influence is an NP hard problem,whose approximate algorithm has low accuracy and complicated computation.In order to solve the above problems,this paper proposes a multi-level inheritance influence and generalization algorithm based on knowledge graph to realize the calculation of inheritance influence and inheritance relationship.The algorithm uses the breadth first search hierarchy computing model of knowledge graph,and takes into account the time span constraints to calculate the inheritance influence and inheritance chain.In order to optimize the computational efficiency,the strategy of depth first search and different levels with different weights is further used to only calculate the influence of the top n levels.The above method can not only calculate the inheritance influence and inheritance chain well,but also can be generalized into various communication influence calculation models.On this basis,this paper proposes a local optimal search similarity algorithm to maximize the propagation influence by selecting the nodes with large propagation influence as spare nodes.It achieves competitive results in running speed and the maximum number of propagation nodes.Finally,the effectiveness of the proposed method is verified by a variety of simulation experiments.

Keywords represent the theme of the text,which is the condensed concept and content of the text.Through keywords,readers can quickly understand the gist and idea of the text and improve the efficiency of information retrieval.In addition,keyword extraction can also provide support for automatic text summarization and text classification.In recent years,research on automatic keyword extraction has attracted wide attention,but how to extract keywords from documents accurately remains a challenge.On the one hand,the keyword is people’s subjective understanding,judging whether a word is a keyword itself is subjective.On the other hand,Chinese words are often rich in semantic information and it is difficult to accurately extract the main idea expressed in the text by solely relying on traditional statistical features and thematic features.Aiming at the problems of low accuracy,information redundancy and information missing in Chinese keyword extraction,this paper proposes an unsupervised keyword extraction method combining knowledge graph and pre-training model.Firstly,topic clustering is carried out by using the pre-training model,and a sentence-based clustering method is proposed to ensure the coverage of the final selected keyword.Then,the knowledge graph is used for entity linking to achieve accurate word segmentation and semantic disambiguation.After that,the semantic word graph is constructed based on the topic information to calculate the semantic weight between words.Finally,keywords are sorted by the weighted PageRank algorithm.Experiments are conducted on two public datasets,DUC 2001 and CSL,and a separate annotated CLTS dataset,the prediction accuracy,recall rate and F1 score are taken as indicators in comparative experiments.Experimental results show that the accuracy of the proposed method has improved compared with other baseline methods,F1 value is increased by 9.14% compared with the traditional statistical method TF-IDF,and increased by 4.82% compared with the traditional graph method TextRank on CLTS dataset.

Patent knowledge graph plays a important role in patent accurate retrieval,patent in-depth analysis and patent know-ledge training.This paper proposes a practical patent knowledge graph construction method based on seed knowledge graph,text mining and relationship completion.In this method,to ensure the quality,a seed patent knowledge graph is first established ma-nually,then the concept and relation extraction method of patent text pattern is used to expand the seed patent knowledge graph,and finally the extended patent knowledge graph is quantitatively evaluated.In this paper,artificial extraction of seed knowledge and manual summarization of lexical and syntactic patterns are carried out for patents in the field of traditional Chinese medicine.After obtaining new lexical and syntactic patterns by machine learning,the knowledge graph of seed patent is expanded and completed.Experimental results show that the number of nodes and relationships in the knowledge graph of traditional Chinese medicine are 19 453 and 194 775 respectively.After expansion,they reach 558 461 and 7 275 958 respectively,representing an increase of 27.7 and 36.3 folds respectively.

Object detection is one of the most popular directions in computer vision,which is widely used in military,medical and other important fields.However,most object detection models can only recognize visible objects.There are often covered(invisible) target objects in pictures in daily life.It is difficult for existing object detection models to show ideal detection performance for covered objects in pictures.Therefore,this paper proposes a novel class reasoning model towards covered area in given image based on informed knowledge graph reasoning and multi-agent collaboration(IMG-KGR-MAC).Specifically,first,IMG-KGR-MAC constructs a global prior knowledge graph according to the visible objects of all pictures in a given picture library and the positional relationship between them.At the same time,according to the objects contained in the pictures themselves and their positional relationships,picture knowledge graphs are established for each picture respectively.The covered objects information in each picture is not included in the global prior knowledge graph and the picture's own knowledge graph.Second,deep deterministic policy gradient(DDPG) deep reinforcement learning idea is adopted to build two cooperative agents.Agent 1 selects the “category label” that is most suitable for the covered object from the global prior knowledge graph according to the semantic information of the current picture,and adds it to the knowledge graph of the given picture as a new entity node.Agent 2 further selects 〈entities,relationships〉 from the global prior knowledge graph according to the newly added entities of agent 1,and expands the graph structure associated with the new entity nodes.Third,agent 1 and agent 2 share the task environment and communicate the reward value,and cooperate with each other to carry out forward and reverse reasoning according to the principles of ‘picture covered target(entity) → associated graph structure' and ‘associated graph structure → picture covered object(entity)',so as to effectively estimate the most likely category label of the covered object of a given picture.Experimental results show that,compared with the existing related methods,the proposed IMG-KGR-MAC model can learn the semantic relationship between the covered picture of a given picture and the global prior knowledge graph,effectively overcome the shortcomings of the existing models that it is difficult to detect the covered object,and has good reasoning ability for the covered object.It has more than 20% improvement in many indicators such as MR(mean rank) and mAP(mean average precision).

Knowledge Graph-to-Text is a new task in the field of knowledge graph,which aims to transform knowledge graph into readable text describing these knowledge.With the deepening of research in recent years,the generation technology of Graph-to-Text has been applied to the fields of product review generation,recommendation explanation generation,paper abstract generation and so on.The translation model in the existing methods adopts the method of first-plan-then-realization,which fails to dynamically adjust the planning according to the generated text and does not track the static content planning,resulting in incohe-rent semantics before and after the text.In order to improve the semantic coherence of generated text,a Graph-to-Text model based on dynamic memory and two-layer reconstruction enhancement is proposed in this paper.Through three stages of static content planning,dynamic content planning and two-layer reconstruction mechanism,this model makes up for the structural difference between knowledge graph and text,focusing on the content of each triple while generating text.Compared with exis-ting generation models,this model not only compensates for the structural differences between knowledge graphs and texts,but also improves the ability to locate key entities,resulting in stronger factual consistency and semantics in the generated texts.In this paper,experiments are conducted on the WebNLG dataset.The results show that,compared with the current exis-ting models in the task of Graph-to-Text,the proposed model generates more accurate content planning.The logic between the sentences of the generated text is more reasonable and the correlation is stronger.The proposed model outperforms existing methods on me-trics such as BLEU,METEOR,ROUGE,CHRF++,etc.

The existing temporal knowledge graph completion models only consider the structural information of the quadruple itself,ignoring the implicit neighbor information and the constraints of relationships on entities,which leads to the poor perfor-mance of the models on the temporal knowledge graph completion task.In addition,some datasets exhibit unbalanced distribution in time,which makes it difficult for model training to achieve a good balance.To address these problems,the paper proposes a context-aware model based on relation constraints(CARC).CARC solves the problem of an unbalanced distribution of datasets in time through an adaptive time granularity aggregation module and uses a neighbor-aggregator to integrate contextual information into entity embeddings to enhance the embedding representation of the entity.In addition,the quadruple relation constraint mo-dule is designed to make the embeddings of entities with the same relational constraints close to each other,while those with diffe-rent relational constraints are far away from each other,which further enhances the embedding representation of entities.Extensive experiments are conducted on several publicly available temporal datasets,and the experimental results prove the superiority of the proposed model.

Redundant code is common in commercial and open source software,and its presence can increase memory footprint,affect code maintainability,and increase maintenance costs.Rapid type analysis algorithm is a common static analysis method in Java redundant code detection,but it still has some shortcomings in virtual method analysis.XTA is a call graph construction algorithm with high precision and efficiency in handling virtual method calls.A method based on XTA call graph construction algorithm is proposed to detect redundant code in Java code.This method is implemented in a prototype tool called redundant code Detection(RCD),and the knowledge graph is constructed to assist manual review to improve the efficiency of manual review and the reliability of redundant code detection.RCD is compared with three other redundant code detection tools by experiments on four open source Java applications.Experimental results show that RCD improves the accuracy of detecting redundant codes by 1%~30% compared with other tools,and improves the integrity of detecting redundant virtual methods by about 4%.

With the expansion of the scope in informatization fields,lots of text corpora in specific fields continue to appear.Due to the impact of security and sensitivity,the text corpora in these specific fields(e.g.,medical records corpora and communication corpora) are often small-scaled.It is difficult for traditional word embedding learning methods to obtain high-quality embeddings on these corpora.On the other hand,there may exist many out-of-vocabulary words in these corpora when using the existing pre-training language models directly,for which,many words cannot be represented as vectors and the performance on downstream tasks are limited.Many researchers start to study how to infer the semantics of out-of-vocabulary words and obtain effective out-of-vocabulary word embeddings based on fine-grained semantic information.However,the current models utilizing fine-grained semantic information mainly focus on the English corpora and they only model the relationship among fine-grained semantic information by simple ways of concatenation or mapping,which leads to a poor model robustness.Aiming at addressing the above problems,this paper first proposes to construct a fine-grained knowledge graph by exploiting Chinese word formation rules,such as the characters contained in Chinese words,as well as the character components and pinyin of Chinese characters.The know-ledge graph not only captures the relationship between Chinese characters and Chinese words,but also represents the multiple and complex relationships between Pinyin and Chinese characters,components and Chinese characters,and other fine-grained semantic information.Next,the relational graph convolution operation is performed on the knowledge graph to model the deeper relationship between fine-grained semantics and word semantics.The method further mines the relationship between fine-grained semantics by the sub-graph readout,so as to effectively infer the semantic information of Chinese out-of-vocabulary words.Experimental results show that our model achieves better performance on specific corpora with a large proportion of out-of-vocabulary words when applying to tasks such as word analogy,word similarity,text classification,and named entity recognition.

In the process of digitisation of medical data,choosing the right technology for efficient processing and accurate analysis of medical data is a common problem faced by the medical field today.The use of knowledge graph technology with the excellent association and reasoning capabilities to process and analyse medical data can better enable applications such as wise information technology of medicine and aided diagnoses.The complete process of constructing a medical knowledge graph includes know-ledge extraction,knowledge fusion and knowledge reasoning.Knowledge extraction can be subdivided into entity extraction,relationship extraction and attribute extraction,while knowledge fusion mainly includes entity alignment and entity disambiguation.Firstly,the constructiontechnologies and practical applications of medical knowledge graphs are summarised,and the development of the technologies is clarified for each specific construction process.On this basis,the relevant techniques are introduced,and their advantages and limitations are explained.Secondly,introducing several medical knowledge graphs that are being successfully applied.Finally,based on the current state of technology and applications of knowledge graphs in the medical field,future research directions for knowledge graphs in technology and applications are given.

Knowledge graphs describe objective knowledge in the real world in a structured form,and are confronted with issues of completeness and newly-added knowledge.As an important means of complementing and updating knowledge graphs,know-ledge graph reasoning aims to infer new knowledge based on existing knowledge.In recent years,the research on knowledge graph reasoning based on repre-sentation learning has received extensive attention.The main idea is to convert the traditional reasoning process into semantic vector calculation based on the distributed representation of entities and relations.It has the advantages of fast calculation efficiency and high reasoning performance.In this paper,we review the knowledge graph reasoning based on repre sentation learning.Firstly,this paper summarizes the symbolic representation,data set,evaluation metric,training method,and evaluation task of knowledge graph reasoning.Secondly,it introduces the typical methods of knowledge graph reasoning,including translational distance and semantic matching methods.Thirdly,multi-source information fusion-based knowledge graph reasoning methods are classified.Then,neural network-based reasoning methods are introduced including convolutional neural network,graph neural network,recurrent neural network,and capsule network.Finally,this paper summarizes and forecasts the future research direction of knowledge graph reasoning.

Due to its rich and intuitive expressivity,knowledge graph has received much attention of many scholars. A lot of works have been accumulated in knowledge graph embedding. The results of the works have played an important role in some fields, such as e-commerce, finance,medicine, transportation and intelligent Q & A. In the knowledge graph embedding model, the loss function plays a key role in its training stage. Based on the existing research of knowledge graph embeddings, this paper classifies the loss functions used in the model into six categories: hinge loss, logistic loss, cross entropy loss, log likelihood loss, negative sampling loss and mean square error loss. The prototype formula and physical meaning of loss functions and their expansion, evolution and application in knowledge graph embedding models are analyzed in detail one by one.Based on the above,the usage, efficiency and convergence of various loss functions in the static and dynamic knowledge graph scenarios are comprehensively analyzed and evaluated. According to the analysis results, combined with the development and application trend of knowledge graph and the current situation of loss functions,the future works of loss functions are discussed.

Knowledge graphs(KGs)has gradually become valuable asset in the field of AI.However,a major problem is that there are many missing edges in the existing KGs.KGs representation learning can effectively solve this problem.The quality of representation learning depends on how well the geometry of the embedding space matches the structure of the data.Euclidean space has been the main force for embeddings;hyperbolic andspherical spaces gaining popularity due to their ability to better embed new types of structured data.However,most data are highly heterogeneous,the single-space modeling leads to large information distortion.To solve this problem,inspired by MuRP model,mixed-curve space model is proposed to provide representations suitable for heterogeneous structural data.Firstly,the Descartes product of Euclidean hyperbolic and spherical spaces is used to construct mixed space.Then,a graph attention mechanism is designed to obtain the importance of relationship.Experimental results on three KGs benchmark datasets show that the proposed model can effectively alleviate the problems caused by heterostructural embedding in low-dimensional spaces with constant curvature.The proposed method is applied to the cold start problem of recommender system,and the corresponding indicators have been improved to a certain extent.

With the development of big data technology,a large amount of heterogeneous data has been generated in various fields.Constructing knowledge graph is an important means to realize semantic intercommunication of heterogeneous data.It is a common method to generate instance model by matching structured data with ontology model mapping.However,most of the existing construction methods require users to manually complete all mapping matching,and the mapping operation is time-consuming and error-prone,unable to perform intelligent matching.In addition,the existing methods do not support incremental updates of the instances.This paper analyzes the existing instance construction methods,and proposes an instance construction and evolution method based on intelligent mapping recommendation to solve the problem of cumbersome manual mapping.Before manually mapping by users,the mapping reuse recommendation mechanism performs multilevel similarity calculation,including element-level similarity,table-level similarity and inter-table propagation similarity,and generates recommendation mapping according to the sorting result of matching.In addition,the incremental discovery mechanism can automatically discover redundant and conflicting instances and generate system background tasks for processing,so as to realize efficient and repeatless import of instances.Experiments are carried out on Shandong government open dataset and Shenzhen medical emergency dataset.With the help of the mapping reuse recommendation module,the interaction time is 3~4 times shorter than that of the traditional mode,and the matching accuracy of field recommendation reaches 98.1%.In the experiment of incremental discovery mechanism,the time required to import 13.94 million instance nodes and 21.58 million relationship edges is reduced from 31.21h to 2.23h,which proves the availability and matching accuracy of intelligent mapping reuse recommendation,and improves the efficiency of instance layer construction and growth.